Two-Stage Graph-Clustering Algorithm and Localized Classification Model to Identify Apt Business Locale
Previous studies conducted in the economy, finance and business management fields have found that there exists a collection of business agglomerations, which contain various numbers of firms that are spread on a specific region. Based on this realness, the selection of apt business locale for a new establishment should then be considered as a trial to identify the prospective business agglomeration in which the new establishment would be able to compete with existing firms. Consequently, a pertinent method that works by characterizing the business agglomerations from a collection of business firms data and subsequently computes the projection of business performance level of a new establishment in each identified agglomeration is developed in this study. A two-stage graph-clustering algorithm that purposively designed to unravel the task of business agglomerations identification is introduced, whereas the localized classification models perform the prediction of business performance level in each known agglomerations. Decisively, results from conducted experiment suggest that the proposed method is beneficial to distinguish the apt business locale for new establishments in a particular region with a collection of business agglomerations.